http://electionarchive.org/ucvAnalysis/US/Exit_Polls_2004_Edison-Mitofsky.pdf

Officially, President Bush won November’s election by 2.5%, yet exit polls showed Kerry winning by 3% [1] . According to a report to be released today by a group of university statisticians, the odds of a discrepancy this large between the national exit poll and election results happening by accident are close to 1 in a million.

In other words, by random chance alone, it could not have happened. But it did.

Two alternatives remain. Either something was wrong with the exit polling, or something was wrong with the vote count.

Exit polls have been used to verify the integrity of elections in the Ukraine, in Latin America, in Germany, and elsewhere. Yet in November 2004, the U.S. exit poll discrepancy was much more than normal exit poll error (and similar to that of the invalid Ukraine election.[2] )

In a recent survey of US members of the world’s oldest and largest computer society, The Association for Computing Machinery, 95% opposed software driven un-auditable voting machines [3] , of the type that now count at least 30% of U.S. votes. Today’s electronic vote-counting machines are not required to include basic safeguards that would prevent and detect machine or human caused errors, be they innocent or deliberate. [4]

The consortium that conducted the presidential exit polls, Edison/Mitofsky, issued a report in January suggesting that the discrepancy between election results and exit polls occurred because Bush voters were more reticent than Kerry voters in response to pollsters.

The authors of this newly released scientific study "Analysis of the 2004 Presidential Election Poll Discrepancies" consider this "reluctant Bush responder" hypothesis to be highly implausible, based on extensive analysis of Edison/Mitofsky’s exit poll data. They conclude, “The required pattern of exit poll participation by Kerry and Bush voters to satisfy the exit poll data defies empirical experience and common sense under any assumed scenario.”

A state-by-state analysis of the discrepancy between exit polls and official election results shows highly improbable skewing of the election results, overwhelmingly biased towards the President.

The report concludes, “ We believe that the absence of any statistically-plausible explanation for the discrepancy between Edison/Mitofsky’s exit poll data and the official presidential vote tally is an unanswered question of vital national importance that needs thorough investigation.”

Ph.D. statisticians in America who have seen this group’s preliminary exit poll study have not refuted it. This new study is a much more comprehensive an analysis of the exit poll discrepancies.

The report is available on-line:

http://electionarchive.org/ucvAnalysis/US/Exit_Polls_2004_Edison-Mitofsky.pdf

An executive summary of the report by is available at:

http://electionarchive.org/ucvAnalysis/US/Exit_Polls_summary.pdf

*Contributors and Supporters of the Report include:*

*Josh Mitteldorf*, PhD - Temple University Statistics Department

*Steven F. Freeman*, PhD - Center for Organizational Dynamics, University of Pennsylvania

*Brian Joiner*, PhD - Prof. of Statistics (ret) University of Wisconsin

*Frank Stenger*, PhD - Professor, School of Computing, University of Utah

*Richard G. Sheehan*, PhD -Professor, Department of Finance, University of Notre Dame

*Paul F. Velleman*, PhD - Associate Prof., Department of Statistical Sciences, Cornell University

*Victoria Lovegren*, PhD - Department of Mathematics, Case Western Reserve University

*Campbell** B. Read*, PhD - Prof. Emeritus, Department of Statistical Science, Southern Methodist University

*Jonathan Simon*, J.D., National Ballot Integrity Project

*Ron Baiman, *PhD* *- Institute of Government and Public Affairs, University of Illinois at Chicago